Imagination Technologies Limited

Daniel Barnard

Clifford Gibson

David Hough

Chris Martin

Colin McQuillan

Gregory Clark

Mike Livesley

Colin Mcquillan

A method for providing input data for a layer of a convolutional neural network (CNN). Input data is received comprising input data values to be processed in a layer of the CNN. Addresses in banked memory of a buffer are determined in which the received data values are to be stored based upon format data indicating a format parameter of the input data in the layer and indicating a format parameter of a filter which is to be used to process the input data in the layer. The received input data values are stored at the determined addresses in the buffer for retrieval for processing in the layer.

Buffer addressing for a convolutional neural network

Shader processing units for a graphics processing unit that are configured to execute one or more ray tracing shaders that generate ray data associated with one or more rays. The ray data for a ray includes a plurality of ray data elements. The shader processing unit comprises local storage, and store logic. The store logic is configured to receive, as part of a ray tracing shader, a ray store instruction that comprises: (i) information identifying a store group of a plurality of store groups, each store group of the plurality of store groups comprising one or more ray data elements of the plurality of ray data elements, and (ii) information identifying one or more ray data elements of the identified store group to be stored in an external unit). In response to receiving the ray store instruction, the store logic retrieves the identified ray data elements for one or more rays from the storage. The store logic then sends one or more store requests to an external unit which cause the external unit to store the identified ray data elements for the one or more rays.

Methods and hardware logic for writing ray tracing data from a shader processing unit of a graphics processing unit

A ray tracing unit and method for processing a ray in a ray tracing system performs intersection testing for the ray by performing one or more intersection testing iterations. Each intersection testing iteration includes: (i) traversing an acceleration structure to identify the nearest intersection of the ray with a primitive that has not been identified as the nearest intersection in any previous intersection testing iterations for the ray; and (ii) if, based on a characteristic of the primitive, a traverse shader is to be executed in respect of the identified intersection: executing the traverse shader in respect of the identified intersection; and if the execution of the traverse shader determines that the ray does not intersect the primitive at the identified intersection, causing another intersection testing iteration to be performed. When the intersection testing for the ray is complete, an output shader is executed to process a result of the intersection testing for the ray.

Intersection testing in a ray tracing system

A system and method for performing intersection testing of rays in a ray tracing system. The ray tracing system uses a hierarchical acceleration structure comprising a plurality of nodes, each identifying one or more elements for intersection testing. The system defines and updates progress information that identifies, for a ray, leaf nodes of the hierarchical acceleration structure which identify elements for which it is not yet known whether or not the ray interests.

Intersection testing for ray tracing

Hardware implementations of, and methods for processing, a convolution layer of a DNN that comprise a plurality of convolution engines wherein the input data and weights are provided to the convolution engines in an order that allows input data and weights read from memory to be used in at least two filter-window calculations performed either by the same convolution engine in successive cycles or by different convolution engines in the same cycle. For example, in some hardware implementations of a convolution layer the convolution engines are configured to process the same weights but different input data each cycle, but the input data for each convolution engine remains the same for at least two cycles so that the convolution engines use the same input data in at least two consecutive cycles.

Hardware implementation of convolutional layer of deep neural network

A system and method for performing intersection testing of rays in a ray tracing system. The ray tracing system uses a hierarchical acceleration structure comprising a plurality of nodes, each identifying one or more elements able to be intersected by a ray. The system makes use of a serial-mode ray intersection process, in which, when a ray intersects a bounding volume, a limited number of new ray requests are generated.

A system and method for performing intersection testing of rays in a ray tracing system. The ray tracing system uses a hierarchical acceleration structure comprising a plurality of nodes, each identifying one or more elements able to be intersected by a ray. The system iteratively obtains ray requests, each of which identifies a ray and a node against which the ray is to be tested, and performs intersection testing based on the ray requests. The number of ray requests obtained in each iteration reduces responsive to an amount of memory occupied by information relating to the rays (undergoing intersection testing) increasing.

Input data for a layer of a convolutional neural network (CNN) is provided by receiving input data values to be processed in a layer of the CNN. Addresses in banked memory of a buffer are determined in which the received data values are to be stored based upon format data indicating a format parameter of the input data in the layer and indicating a format parameter of a filter which is to be used to process the input data in the layer. The received input data values are stored at the determined addresses in the buffer for retrieval for processing in the layer.

Data for layers of a convolutional neural network (CNN) is provided by receiving input data values to be processed in a layer of the CNN and determining addresses in banked memory of a buffer in which the received data values are to be stored based upon format data indicating a format parameter of the input data in the layer and indicating a format parameter of a filter which is to be used to process the input data in the layer. The received input data values are then stored at the determined addresses in the buffer for retrieval for processing in the layer.

Growing community of inventors

Berkamsted, United Kingdom

Daniel Barnard